Janhunen J. The Mongolic Languages

Подождите немного. Документ загружается.

Consonant stems were subdivided into obstruent stems, ending in the stops *b *d *g

or the sibilant *s, and sonorant stems, ending in the nasals *m *n *ng or the liquid *l.

This division correlates with the alternation of weak and strong obstruents in certain

suffixes, as in *ol- ‘to find’ : pass. *ol.da- vs. *ab- ‘to take’ : pass. *ab.ta-. Rather unex-

pectedly, the functional obstruent stems also comprised the stems ending in the liquid

*r, as in *ger ‘dwelling’ : dat. *ger-tü/r vs. *gal ‘fire’ : dat. *gal-du/r. This peculiarity,

preserved in most Modern Mongolic languages, suggests that *r, at least word-finally,

may originally (in Pre-Proto-Mongolic) have been a true obstruent. It is true, *r seems

also to have had a functional affinity with the other liquid *l, for neither of the two liq-

uids was originally permitted in word-initial position. The only other consonant with this

restriction was *ng. There are, however, several Common Mongolic words beginning

with *l, e.g. *luu ‘dragon’ (from Chinese, through Uighur).

WORD FORMATION

Apart from vowel harmony and the insertion of connective segments at the border of

stem and suffix, Proto-Mongolic morphology was based on a rather mechanical aggluti-

nation of derivative and inflectional suffixes to essentially invariable stems. There were

two major parts of speech which may be identified as nouns (nominals) and verbs

(verbals), combined with two separate sets of suffixes, respectively. Morphological and

syntactic details allow nouns to be further divided into substantives, pronouns, and

numerals. Some nominal (including pronominal) stems, often with a defective or excep-

tional paradigm, functioned as adverbs and postpositions. Adjectival words were also

basically nominal, though their derivatives could function as verbs, cf. e.g. *ca.ga.xan

‘white’ : ess. *ca.yi- ‘to be white’, *köke ‘blue’ : transl. *köke.re- ‘to become blue’.

Nominal and verbal stems had a basically identical structure, and some stems (nomina-

verba) can actually be reconstructed as having had both a nominal and a verbal function,

e.g. *emkü- ‘to put into mouth’: *emkü ‘bite’. Such cases could perhaps be analysed as

examples of zero derivation, but synchronically it is impossible to determine which of the

two functions (nominal or verbal) should be viewed as derivationally primary. Both nom-

inal and verbal stems could end in a vowel, the liquids *l *r, or any of the obstruents *b

*d *s *g. Importantly, however, there were no verbal stems ending in a nasal, while all the

three nasals *m *n *ng are well attested as the final segments of nominal stems.

The nominal stems ending in the nasal *n may be viewed as a separate stem type,

perhaps best identified as the nasal stems (proper). There were two kinds of nasal stem:

those ending in a morphophonologically stable *n and those ending in an unstable or

‘fleeting’ */n. The unstable */n was in a regular paradigmatic alternation with zero (Ø),

as in *mori/n ‘horse’ :gen. *morin-u : acc. *mori/y-i. The fact that the unstable */n was

not permanently present in the stem suggests that it may originally have been a suffix. Its

original function remains, however, unclear; it may have been a derivative suffix,

perhaps denoting a specific class of nouns, but it may also have been connected with the

categories of number and case. It is probably relevant to note that the stems ending in the

unstable */n were much more numerous than those ending in the stable *

n, a situation

that is still valid for many (though not all) Modern Mongolic languages.

In accordance with the two basic parts of speech in the language, the derivative

suffixes that can be reconstructed for Proto-Mongolic may be divided into four types,

depending on whether they produced (1) denominal nouns, (2) denominal verbs,

(3) deverbal nouns, or (4) deverbal verbs. Each type of derivative had a specific set of

10 THE MONGOLIC LANGUAGES

suffixes, many of which are still productive in the Modern Mongolic languages. From

the structural point of view, the denominal derivative suffixes are relatively uninterest-

ing, though some of them seem to have been extremely productive, such as *.tU or

*.tA.( y)i [possessive adjectival nouns], *.lA- [denominal verbs with a variety of

functions]. Among the more restricted and less commonplace categories of derivation

was gender (female sex), indicated by the denominal suffixes *.jin [female beings, from

tribal names and age expressions] and *.gcin [female animals, from colour terms].

A higher degree of grammaticalization was present in the deverbal verbs, most of

which may be understood as expressions of the category of voice, comprising the sub-

classes of passive, causative, reciprocative, cooperative, and pluritative verbs. Passives

were marked by the suffix variants *.dA- (after sonorant stems), *.tA- (after obstruent

stems), and *.g.dA- (after vowel stems); causatives by *.gA- (after sonorant stems and

stems in *r), *.kA- (after obstruent stems), and *.xA-, *.l.gA- or *.xUl- (after vowel

stems); reciprocals by */U.ldU- ; cooperatives by */U.lcA- ; and pluritatives by *.cAgA- .

The details of the actual formation of these derivatives were already in Proto-Mongolic

to some extent lexicalized. Some stems had, for instance, two alternative causatives, as

in *bayi- ‘to be’ : caus. *bayi.xul- or *bayi.lga-. There were also double causatives, as

in *gar- ‘to exit’ : *gar.ga.xul- ‘to cause to take out’.

Deverbal nouns were likewise inherently liable to be grammaticalized, and it is in

some cases difficult to draw a distinction between derivational deverbal nouns and the

inflectional category of participles (verbal nouns). The basic criteria are the degree of

productivity and verbality of the derived nominal stems. Participles may be defined as

fully productive deverbal nouns, which still function as verbal headwords in the sen-

tence. In Proto-Mongolic there were, however, many cases of lexicalized participles

which had apparently lost their verbal characteristics (or never developed them), like

*ide- ‘to eat’ : *ide.xe/n ‘food’ vs. part. imperf. *ide-xe. On the other hand, some Proto-

Mongolic and/or Common Mongolic deverbal nouns, like those in *.l (general action)

and *.mAr (potential action), function very much like participles, but are, nevertheless,

in Mongolic studies normally counted as derivational deverbal nouns.

One of the most difficult borderline cases involves the actor nouns or agentive

participles in *-g.ci or *-xA.ci. Normally listed as participles, these forms seem, indeed,

to have had a number of verbal functions in Proto-Mongolic. Many actual examples of

actor nouns are, however, better analysed as lexicalized regular nouns, like *jiru.g.ci or

*jiru.xA.ci ‘artist’, from *jiru- ‘to draw, to paint’. In most Modern Mongolic languages,

the verbal features of the actor nouns are absent or very marginal. It may also be noted

that of the two alternative suffix variants, the variant *-xA.ci is based on the imperfective

participle suffix *-xA, as in part. imperf. *jiru-xA ‘(the act of) drawing’, while the

variant *-g.ci

is a secondary derivative of the non-productive deverbal noun in *.g, as in

*jiru.g ‘picture’.

Most of the Proto-Mongolic suffixes for deverbal nouns yielded clearly nominal

formations with restricted productivity and a tendency of lexicalization. The derivatives

concerned may be characterized as various types of general action nouns, such as those

in *.dAl, *.lAng, *.lgA, *.ltA, *.li, *.m, *.mji, *.mtA, *.r. Some were, however, more

specialized and yielded nouns denoting, for instance, place of action: *.ri, *.xUri or

*.xUli; result or object of action: *.jA, *.mAg, *.ng, *.si; state or quality resulting from

action: *.xU or *.xUn, *.gAyi or *.gAr, *.mAl ; performer of action: *.xUl ; or instrument

of action: *.xUr. It is easy to see that many of these suffixes contain certain recurrent

initial elements, e.g. *.l, *.m, *.xU, which may be identified as their original primary

components, to which additional elements were added secondarily.

PROTO-MONGOLIC 11

In deviation from the general dominance of suffixal derivation, there is a single aber-

rant phenomenon, in which prefixation also plays a role. This is the Common Mongolic

pattern of forming emphatic (intensive) derivatives from adjectival nouns, especially

colour terms, by prepositing to the stem its partially reduplicated initial syllable followed

by the consonant .b according to the formula (C)V.b&(C)V-, e.g. *xulaxan ‘red’ :

*xu.b&xulaxan ‘reddish, quite red’. The reduplicated syllables may in such cases be

analysed as independent emphatic particles, but in some modern languages they have

yielded fully lexicalized structures, as in Bonan shera ‘yellow’ (< *sira) : shew.rexang

‘quite yellow’ (< *si.b+sira.kan).

A different type of reduplication is involved in the formation of the generic plural

(‘and other things like that’). Already in Proto-Mongolic, the generic plural seems to

have been formed by pairing the nominal stem with an echo word, which was either

a rhyme beginning with *m or an alternate containing *a in the initial syllable, e.g.

*noka( y)i & moka( y)i ‘dogs and the like’, *mori & mari ‘horses and the like’. In spite

of its marginal function the generic plural, with some variations in the actual patterns, is

surprisingly widely attested throughout the Mongolic family.

NOMINAL NUMBER

The morphological categories characteristic of the nominal declension in Proto-

Mongolic were number, case, and reflexive possession. Unlike case and reflexive pos-

session, however, number was not a regular inflexional category, but rather a derivational

feature involving a considerable degree of facultativeness and irregular lexicalized vari-

ation. This need not always have been so, for there are indications that number marking

had undergone a secondary diversification in Pre-Proto-Mongolic. This diversification

has continued in some Modern Mongolic languages, while in others a strictly limited set

of inflexional number suffixes has been established. The marked number in Mongolic

has always been the plural (collective), but in some stem types the plural markers replace

elements that may originally have functioned as singular (singulative) suffixes.

The plural in Proto-Mongolic was marked by two basic suffixes, *.s and *.d, which

were in complementary distribution. The suffix *.s was added to vowel stems, e.g. *ere

‘man’ : pl. *ere.s, while the suffix *.d, preceded by the connective vowel *U (> *UU),

was added to consonant stems, e.g. *nom ‘book’ : pl. *nom/u.d. However, most stems

ending in the consonants *n *l *r lost the final segment before the suffix *.d, with no

connective vowel involved, e.g. *kan ‘prince’ : pl. *ka.d, suggesting that these final con-

sonants may originally have been suffixes. This analysis is particularly likely in the case

of the stems ending in the unstable */n, which regularly formed their plural by the suffix

*.d, e.g. *mori/n ‘horse’ : pl. *mori.d. The same is true of polysyllabic stems ending in

the derivative complex *.sU/n, e.g. *nugu.su/n ‘duck’ : pl. *nugu.d.

A third plural suffix, with a more restricted distribution, was *.n, which regularly

replaced the stem-final derivative element *.( y)i (possibly < *.xi) in several complex suf-

fixes, e.g. poss. *.tA.( y)i : pl. *.tA.n, part. fut. *-kU.( y)i : pl. *-kU.n. This suffix was also

used with the actor noun marker part. ag. *-g.ci : pl. *-g.ci.n. Occasional traces of *.n are

still preserved in the Modern Mongolic languages, but generally it has lost its produc-

tivity in favour of the other plural suffixes. In Common Mongolic, regular nouns ending

in *.( y)i > *.i form their plural by the suffix *.s, e.g. *noka.i ‘dog’ : pl. *noka.s, but

evidence from Middle Mongol and Preclassical Written Mongol shows that the original

pattern is likely to have involved the use of *.n, i.e. *noka.i : pl. *noka.n.

12 THE MONGOLIC LANGUAGES

Already in Proto-Mongolic, the basic plural suffixes were being complemented by a

set of secondary suffixes. Some of the latter were simply semantically redundant

reduplications or combinations of the basic suffixes (double plurals), e.g. *.d/U.d

(> *.dUUd ), *.s/U.d (> *.sUUd ). Others may be analysed as combinations of original

stem-final segments or syllables with the plural formative */U.d, e.g. Common Mongolic

*.nUUd, *.ciUd (> *.ciUl). A different type of innovation was involved in the element

*.nAr, also *.nA.d or *.nar/U.d (> *.nar.UUd), which was added to nouns denoting

humans or deities, e.g. *aka ‘elder brother’ : pl. *aka.nar, *tenggeri ‘god’ : pl.

*tenggeri.ner. Owing to the diversification of the plural suffixes, the original rules of

complementarity were lost, often allowing several different plurals to be formed of a

single nominal stem.

Patterns of the type *mori/n (*mori.n) : *mori.d and *noka.(y)i : *noka.n suggest that

plural formation may originally have been part of a more general system of nominal

classes, in which both the singular and the plural were marked by distinct class suffixes.

What the semantic basis of this possible Pre-Proto-Mongolic class system may have

been, remains to be clarified, but in any case it is obvious that the distribution of the

plural suffixes was not only phonologically conditioned. It is unclear what the exact

function of the plural originally was. As in the Modern Mongolic languages, the basic

(singular) form of nouns in Proto-Mongolic was able to function as an unmarked

(unspecified) plural. The use of the actual plural suffixes seems to have been limited to

cases in which plurality was not otherwise obvious from the context.

An interesting perspective into the prehistory of class marking in Mongolic is offered

by the bisyllabic stems ending in *sU/n and *dU/n. The former typically denote liquids

or liquifiable masses: *usu/n ‘water’, *üsü/n ‘milk’, *casu/n ‘snow’, *cisu/n ‘blood’,

*nisu/n ‘mucus’, *tosu/n ‘oil’, while the latter denote countable sets of identical objects:

*modu/n ‘tree/s’, *nidü/n ‘eye/s’, *sidü/n ‘tooth/teeth’, *sodu/n ‘quill feather/s’, *xodu/n

‘star/s’, *xödü/n ‘feather/s’. Simple internal reconstruction suggests that all of these

stems were originally composed of a monosyllabic root (CV), to which a class suffix (*.d

or *.s) was added, followed by the connective vowel *U and the suffixally used unstable

*/n. This system of classes was obscured already in Pre-Proto-Mongolic, but it is perhaps

relevant to note that the two class markers are identical with the two basic plural suffixes

(*.d and *.s) still used in Proto-Mongolic.

NOMINAL CASE

The category of case in Proto-Mongolic is normally considered to have comprised six

suffixally marked cases: genitive, accusative, dative, ablative, instrumental, and comita-

tive. At the Common Mongolic level there are also other case-like forms. The unmarked

basic stem may be regarded as a nominative. The case endings were identical for all

nouns except for slight phonologically conditioned variation depending on the stem type.

If we take vowel stems (V) as the basis, some case endings, though not all, had separate

variants used with consonant stems (C) or, more specifically, with nasal stems (N) or

obstruent stems (O). The case endings were also affected by vowel harmony (Table 1.3).

Just how the actual shapes of the case endings are to be reconstructed depends on

what level of reconstruction is intended. For some details, both Written Mongol and

Middle Mongol yield information that is not readily recoverable from the synchronic

data of the Modern Mongolic languages. Case endings are, in fact, a good example of a

morphological set that should be viewed at three different levels of reconstruction:

Common Mongolic, Proto-Mongolic, and Pre-Proto-Mongolic. At the Pre-Proto-Mongolic

PROTO-MONGOLIC 13

level, the synchronic allomorphy of the Proto-Mongolic case endings can be shown to

derive from original invariance, while at the Common Mongolic level several new

complications had appeared.

A very simple type of allomorphy is involved in the accusative case ending, which

clearly derives from Pre-Proto-Mongolic *-i with the only complication that vowel stems

required the presence of the connective consonant *y. The genitive ending, on the other

hand, may be reconstructed as *-n, which after consonant stems required the connective

vowel *U. After nasal stems, the actual case ending was dropped, leaving only the

connective vowel to signal its former presence: */U-n > *-U. Vowel stems probably

originally took the primary genitive ending *-n, but this was secondarily expanded into

*/y-i-n on the analogy of the consonant stems, and under the influence of the accusative

ending */y-i.

As far as their functions are concerned, the genitive and the accusative may be iden-

tified as the basic grammatical cases in Mongolic, with the genitive marking the adnom-

inal (attributive) and the accusative the adverbal (objective) type of dependence. It has

to be noted that these two cases, although clearly distinct for all stem types in both Proto-

Mongolic and Pre-Proto-Mongolic, show a secondary tendency to merge in several

Modern Mongolic languages, notably Dagur and the languages of the Gansu-Qinghai

complex. From the point of synchronic description, it seems that the formally syncretized

genitive-accusative, combining the functions of its two ancestors, cannot be treated as

two separate cases. Rather, it is a single new case, which is perhaps best termed the

connective.

The history of the dative ending is connected with several unsolved problems. The

full ending *-dUr (*-DUr) is only attested in Written Mongol and Middle Mongol, while

all the Modern Mongolic languages point to the shape *-dU (*-DU). This apparently

means that the ending was irregularly shortened already in Proto-Mongolic. However,

there are indications that the original ending may have been simply *-d, as still attested

in a number of adverbial and postpositional words, such as *uri-d ‘before’. If this is so,

the complex ending *-dUr is best explained as a combination of the elements *d and *r,

joined with the intermediation of the connective vowel *U. The role of the final element

*r remains unclear, though it has been compared with the adverbial suffix *.xUr > *g/-UUr,

which functions as a prosecutive ending (‘via’) in a number of Modern Mongolic lan-

guages. The prosecutive might, however, also be connected with the directive in *-rUU ~

*-UUr, which derives from the independent postposition *uruxu > *uruu ‘down/wards;

towards’ (cf. also *uru-gsi ‘forward’).

In addition to the dative in *-d/U-r > *-dU/r, Proto-Mongolic still had traces of another

case in *-A, often also identified as a dative but perhaps better termed the locative.

14 THE MONGOLIC LANGUAGES

TABLE 1.3 PROTO-MONGOLIC CASE MARKERS

VCNO

gen. */y-i-n */U-n *-U

acc. */y-i *-i

dat. *-dU/r *-tU/r

abl. *-A-cA

instr. *-xAr */i-xAr

com. *-lUx-A

The locative ending is well attested in both Written Mongol and Middle Mongol, but

unlike the dative ending it had the restriction of being added only to consonant stems,

e.g. *gajar ‘place’ : loc. *gajar-a. In Common Mongolic it is mainly preserved in adver-

bial and postpositional items, in which it is often preceded by the elements *.r or *.n, as

in *dexe.r-e ‘on top of’, *emü.n-e ‘in front of’. Even more importantly, the locative end-

ing occurred in combination with the dative element *-d-, yielding *-d-A (*-D-A), which

is attested as an alternative dative ending in Middle Mongol, and in a few Common

Mongolic fixed phrases, e.g. *nasu/n ‘age’ : *nasu-d-a ‘always’. All of this suggests that

the element *-d- of the dative may originally have been a coaffix, to which other

elements were added, yielding the complex dative endings *-d-A and *-d/U-r > *-dU.

Due to the effect of vowel reduction, the endings *-d-A and *-d-U are largely indistin-

guishable in the Modern Mongolic languages.

At the Proto-Mongolic level, the functions of the locative in *-A seem to have been

identical with those of the dative in *-d/U-r, which explains the ultimate marginalization

of the locative. Both cases are attested in a wide range of locative and dative (dative-

locative) functions, expressing not only spatiality (‘where’, ‘whither’) and temporality

(‘when’), but also the recipient (‘for whom’) and possessor (‘in whose possession’).

Analogous functions were filled by the ablative with the difference that it indicated the

source of action (‘from where’, ‘from whom’). The original ablative ending seems to

have been *-cA, still preserved in relicts in Written Mongol, as in vguvca for *exün-ce

‘from this’. Already in Proto-Mongolic, however, the ablative was mainly expressed by

the complex suffix *-A-cA, which incorporates the locative ending *-A-.

The dative and the ablative, together with the remains of the locative, may be identi-

fied as the local cases of Proto-Mongolic. The instrumental and the comitative, corre-

spondingly, were the modal cases, expressing, roughly, the means of action (‘by what’)

and the social context of action (‘with whom’), respectively. The instrumental ending

may be derived from the basic shape *-xAr < *-pAr, expanded into */i-xAr after conso-

nant stems. The comitative ending *-lUx-A incorporates the locative in *-A, revealing

that the comitative was a secondary case formed relatively late in Pre-Proto-Mongolic on

the basis of a denominal derivative suffix for possessive adjectival nouns. This develop-

ment has later recurred, in that the original comitative in the Modern Mongolic languages

has largely been replaced by what may be termed the possessive case, based on the

Common Mongolic possessive adjectival suffix *.tA.(y)i.

In the Post-Proto-Mongolic period, none of the original case endings has developed

along completely regular phonological lines. One particularly conspicuous tendency,

which may be regarded as Common Mongolic, is the secondary lengthening of the

suffix-initial (morpheme-boundary) vowel elements for all stem types. As a result, the

modern genitive and accusative endings typically incorporate the long vowel elements

*-Ai- or *-ii-, while the ablative and instrumental have *-AA-. The vowel element has in

some cases become morphologically distinctive, cf. e.g. *xaan ‘emperor’ : gen. *xaan-ai

vs. acc. *xaan-ii. Another irregular feature is the development *-cA > *-sA in the ablative

ending *-A-cA

, yielding Common Mongolic *-AA-sA. The declensional patterns of the

stems ending in the unstable /n vary considerably in the modern languages.

As the history of the ablative ending *-A-cA shows, Proto-Mongolic had a tendency

to accumulate certain case suffixes to sequences in a pattern that has been termed

double declension. Apart from the occasional combination of two local case endings

(normally dative + ablative), it is particularly common in several Modern Mongolic

languages to form a secondary case paradigm on the basis of the genitive (especially

genitive + dative or ablative). Although the actual forms cannot necessarily be derived from

PROTO-MONGOLIC 15

Proto-Mongolic, the tendency of double declension itself may well be characterized as at

least Common Mongolic.

There is, incidentally, one type of double declension that is definitely both Common

Mongolic and Proto-Mongolic. This involves the use of the suffix *-ki (or *-ki/n) after

the locative, dative and genitive endings to form new nominatives, which can, in principle,

be further inflected in different case forms. The dative ending used in this connection is

always *-dA-(*-DA-), e.g. *ger ‘house’ : *ger-te-ki ‘(the one) located in the house’ : dat.

*ger-te-ki-dü/r. The suffix *-ki is traditionally treated as a derivational feature, but it

differs from all other derivative suffixes in that it is normally added to an inflected form.

Unlike most actual derivative suffixes, it is also fully productive, and must have been so

in Proto-Mongolic already. It is therefore probably best analysed as a special kind of

nominative case ending, used in double declension to nominativize other case forms. The

underlying structural motivation is obvious: the nominative is unmarked in its normal

use, but when built upon other case forms, it is a marked feature and requires an ending,

which is *-ki.

Additional information on the history of the case system is offered by the reflexive

(reflexive-possessive) declension, in which the case endings are followed by the reflex-

ive marker *-xA/n < *-pA/n, after consonants */i-xA/n. The adding of the reflexive marker

seems originally to have been fairly mechanical, though secondary irregularities are

observed in some modern languages especially in the genitive and accusative. Perhaps

most importantly, the dative ending used in the reflexive declension has always been

*-dA (*-DA), yielding the complex *-dA-xA/n (*-DA-xA/n). The reflexive marker could

also follow the bare stem, yielding an unmarked form functionally equivalent to the

accusative.

NUMERALS

In view of its relatively shallow dating, it is not surprising that Proto-Mongolic had a

fully developed set of native numerals, corresponding to a decimal system of counting.

It is, indeed, perhaps more surprising that some of the peripheral Mongolic languages,

notably (Minhe) Mangghuer and Moghol, have replaced the original numeral set, or large

sections of it, by recent borrowings and other innovations. Although this is mainly

indicative of the exceptionally strong areal reorientation of the languages concerned, the

possibility of similar replacements in Pre-Proto-Mongolic should not be overlooked. As

it is, the Mongolic numerals are a promising object for internal reconstruction.

The Proto-Mongolic numerals of the first decade may be reconstructed as: 1 *nike/n >

*nige/n (> Common Mongolic *nege/n), 2 *koxar ~ *koyar, 3 *gurba/n, 4 *dörbe/n,

5 *tabu/n, 6 *jirguxa/n, 7 *doluxa/n, 8 *na( y)ima/n, 9 *yersü/n (> Common Mongolic

*yesü/n), 10 *xarba/n. The other decades were expressed by separate correlative deriva-

tives: 20 *kori/n, 30 *guci/n, 40 *döci/n, 50 *tabi/n, 60 *jira/n, 70 *dala/n, 80 *naya/n,

90 *yere/n. There were also words for the lower powers of ten: 100 *jaxu/n, 1,000

*mingga/n, 10,000

*tüme/n (generically also ‘myriad’).

An examination of the numeral material immediately reveals some diachronically

relevant regularities and irregularities. Most importantly, it may be observed that all

numerals, with a single exception, belong to the same stem type, ending in the unstable

*/n. The exception is 2 *koxar ~ *koyar, which, because of its aberrant shape, is likely to

be a secondary innovation. In fact, it is commonly assumed that the original numeral for

‘two’ was *jiri/n, still used in Middle Mongol for counting female beings. The primary

16 THE MONGOLIC LANGUAGES

status of this stem is confirmed by the fact that the numeral 6 *jirguxa/n is transparently

a compound word, analysable as *jir;guxa/n ‘2 x 3’, with 3 *gu(r)ba/n as the latter

component. This, on the other hand, suggests that the Proto-Mongolic numeral for ‘six’

was also an innovation replacing a more original stem.

Another detail revealed by internal reconstruction is that several basic numerals

contain a derivative suffix which can be reconstructed as Pre-Proto-Mongolic *.pA/n,

yielding Proto-Mongolic *.bA/n : *.mA/n : *.xA/n. The variant *.bA/n occurs in three

numerals after the consonant *r, which itself is also likely to be a derivative suffix:

3 *gu.r.ba/n, 4 *dö.r.be/n, and 10 *xa.r.ba/n. It may be noted that *r also appears in

2 *ji.r.i/n, though the segmentation of this numeral is problematic. The variant *.xA/n

occurs after a vowel in 6 *ji.r+gu.xa/n and 7 *dolu.xa/n, while the variant *.mA/n was

obviously conditioned by the initial nasal in 8 *na( y)i.ma/n. Altogether, *.pA/n was

clearly a suffix making fully formed numerals of the first decade out of a set of abstract

(primary) numeral roots. The absence of *.pA/n in 1 *nike/n > *nige/n, 5 *tabu/n, and

9 *yersü/n suggests that these numerals were somehow special and perhaps secondary.

Further conclusions can be made from the comparison of the basic numerals with the

corresponding set for the decades. The numerals for the decades are clearly divided into

two groups, the first ending in *.i/n, as seen in the items for 20 to 50, and the second end-

ing in *.A/n, as seen in in the items for 60 to 90. The conclusion lies close at hand that

these suffixes represent the meaning ‘ten’, though an immediate comparison with Proto-

Mongolic 10 *xa.r.ba/n appears phonologically impossible. In any case, the elements

preceding *.i/n and *.A/n may be identified as the original roots for the basic numerals

of the first decade, which may then be reconstructed as: 3 *gu(-), 4 *dö(-), 5 *tab, 7 *dal,

8 *nay, 9 *yer. The root in 60 *

jir.a/n represents, of course, 2 *ji.r.

There still remain many unanswered questions about the Mongolic numerals. There

is, for instance, not sufficient internal evidence to explain the alternation *r : *c in

3 *gu.r.ba/n : 30 *gu.c.i/n and 4 *dö.r.be/n : 40 *dö.c.i/n. In 7 *dol/u.xa/n : 70 *dal.a/n

the correspondence *o : *a is easily explained by assuming a sporadic assimilation in

*dol/u.xa/n < *dal/u.xa/n, but it is not clear why the suffix *.xA/n < *.pA/n is here pre-

ceded by what appears to be the connective vowel *U. The root in 8 *na( y)i.ma/n :

80 *nay.a/n is potentially important for the reconstruction of Proto-Mongolic diphthon-

goids, but it is also possible that 8 *na( y)i.ma/n should be segmented as *nay/i.ma/n,

with *i functioning as a connective vowel. In the latter case, the numeral root *nay could

ultimately derive from Pre-Proto-Mongolic *nax (or even *nap).

A few of the original numeral roots are also attested in a limited set of archaic

ordinals, comprising: *ji.tüxer ‘second (wife)’, *gu.taxar ‘third’, *dö.tüxer ‘fourth’,

*tab.taxar ‘fifth’. The ordinal suffix in question shows irregular variation both in the

vocalism (*U : *A) and in the consonantism (*x : *g), but it seems to be based on the

more primary ordinal suffix *.tU or *.tA, expanded by the instrumental case ending

*-xAr. In Common Mongolic, *.tA forms multiplicatives, while the ordinal suffix appears

in the generalized shape *.dUgAr > *.dUgAAr, normally added to the full stems of the

basic numerals, with only the unstable */n omitted, e.g. *

nige.düger ‘first’, *gurba.dugar

‘third’. Occasional irregular truncation of the stem is, however, observed in

*koya(r).dugar ‘second’, *jirgu(xa).dugar ‘sixth’, *dolu(xa).dugar ‘seventh’.

Two other widespread categories of numeral derivative that can unambiguously be

dated back to Proto-Mongolic are the collectives in *.xUlA/n, e.g. *koya.xula/n ‘two

together’, and the distributives (later also approximatives) in *.xAd, e.g. *gurba.xad

‘three each’. The distributive *kosiya.xad ‘two each’ contains an exceptional root variant,

revealing a derivative connection with *kos ‘pair’ (which is probably etymologically

PROTO-MONGOLIC 17

separate from the numeral *koxar ~ *koyar ‘two’). The distributive *niji.xed ‘one each’

is also exceptional. Moreover, *niji.xed is alternatively attested as *niji.xel, suggesting

that the distributive suffix *.xAd was originally a plural (*.xA.d ) from the primary

suffix *.xAl.

PRONOUNS

A major formal difference between pronouns and substantival nouns in Proto-Mongolic

was that the former almost invariably involved maximally simple monosyllabic roots

(CV, in the modern languages often lengthened into CVV). In actual use, however, most

pronominal roots were expanded by derivative and inflectional elements, many of which

were unknown in regular nominal morphology. From the point of view of function, the

three principal categories of pronoun were: personal, demonstrative, and interrogative

pronouns.

The personal pronouns (Table 1.4) formed an almost perfect grid, in which person

was marked by the initial consonant (1p. *b : *m, 2p. *c : *t, 3p. *Ø), and number by the

stem vowel (sg. *i : pl. *a). In the inflected forms, the singular pronouns had two expan-

sions, one for the genitive (*.n-) and the other for the rest of the oblique cases (*.mA-),

while the plural pronouns had only one expansion for the whole paradigm (*.n-). The

system that can be reconstructed for Pre-Proto-Mongolic was even more regular, in that

the alternations of the initial consonants in the first and second person stems can be

derived from original invariance (1p. *m < *b by nasalization, 2p. *c < *t by palatalization).

The same is true of the seemingly irregular first person singular oblique stem (*na.ma- <

*nï.ma- < *mï.ma- < *bï.ma- by nasalization, dissimilation, and prebreaking).

One detail that is not immediately clear from the comparative material concerns the

harmonic status of the singular pronouns 1p. *bi, 2p. *ci, 3p. *i. The fact that the oblique

stems, 1p. *na.ma-, 2p. *ci.ma-, 3p. *i.ma-, are clearly back-vocalic suggests that the

pronominal roots originally contained the velar vowel *ï. However, the velar vocalism of

the oblique stems can also be due to the regressive influence of the element *.mA, which

is etymologically obscure, but which may originally have been back-vocalic itself. It

happens that the corresponding genitive forms are harmonically ambivalent, and have

been variously reconstructed either as back-vocalic, 1p. *mi.n-u, 2p. *ci.n-u, 3p. *i.n-u,

or as front-vocalic, 1p. *mi.n-ü, 2p. *ci.n-ü, 3p. *i.n-ü. Strictly speaking, the genitives

were restructured into 1p. *mi.n-i-

, 2p. *ci.n-i-, 3p. *i.n-i- already in Proto-Mongolic, as

suggested by all the Modern Mongolic languages.

18 THE MONGOLIC LANGUAGES

TABLE 1.4 PROTO-MONGOLIC PERSONAL PRONOUNS

1p. 2p. 3p.

sg. nom. *bi *ci *i

gen. *mi.n-U *ci.n-U *i.n-U

obl. *na.ma- *ci.ma- *i.ma-

excl. incl.

pl. nom. *ba *bida *ta *a

gen. *ma.n-u *bida.n-u *ta.n-u *a.n-u

obl. *ma.n- *bida.n- *ta.n- *a.n-

The reconstruction of the genitives of the personal pronouns is further complicated by

the presence of a set of possessive pronouns, formed from the basic genitives by the

suffix *-xA( y)i. This suffix is functionally close to the general nominativizing element

(double declension nominative ending) *-ki, which is, in fact, also used to form possessive

pronouns in several Modern Mongolic languages. The suffix *-xA( y)i seems to survive in

the modern languages in, at least, the genitives of the plural personal pronouns 1p. *man-

u-xa( y)i > *man-ai, 2p. *tan-u-xa( y)i > *tan-ai. The latter are, however, not confined to

the possessive (predicative) function, but are also used as regular (attributive) genitives. It

is unclear whether a similar merger of the two forms has taken place in the singular.

In the Post-Proto-Mongolic period, the system of personal pronouns has been affect-

ed by three major structural innovations. The first innovation, which was apparently ini-

tiated already in Late Pre-Proto-Mongolic, was the appearance of a secondary inclusive

pronoun for the first person plural. The new pronoun was of a compound origin, con-

sisting of sg. 1p. *bi and pl. 2p. *ta, i.e. *bi+ta ‘I and you’ > *bida (also > *bide). With

this innovation, the original pronoun *ba : *man- was restricted to the exclusive function

and was gradually marginalized. Although the distinction between the inclusive and

exclusive forms is preserved in several Modern Mongolic languages in the oblique par-

adigm, there has been a tendency to merge the categories in the nominative by replacing

*ba by *bida. The only modern language preserving the original set is Dagur.

The second innovation concerned the third person pronouns sg. *i : pl. *a, which

already in Proto-Mongolic were becoming obsolete, and which in Common Mongolic

have been completely replaced by the demonstratives. The original pronouns are still

attested in Middle Mongol and Preclassical Written Mongol, but the only modern

languages preserving them, or traces of them, are Dagur and Moghol.

The third innovation was the honorific use of the plural second person pronoun *ta in

reference to a single person: ‘you [single, honoured one]’. In order to make the plural ref-

erence unambiguous, several Modern Mongolic languages have introduced suffixally

marked plurals, such as *ta.nar (> Common Mongolic *taa.nar) ‘you [many]’. Similar

plurals are also formed of the first person pronoun: *bida.nar (>*bide.ner). While these

innovations serve obvious communicative functions, they have seriously altered the

formal structure of the pronominal system.

Apart from their normal independent use, the personal pronouns in Proto-Mongolic

were apparently used enclitically, which in several Modern Mongolic languages has

resulted in grammaticalized systems of possessive suffixes and predicative personal end-

ings. The possessive suffixes are based on the genitives: 1p. sg. *-mini : pl. excl. *-mani :

incl. *-bidAni, 2p. sg. *-cini : pl. *-tani, 3p. sg. *-ini : pl. *-ani. The predicative personal

endings, on the other hand, are based on the nominatives: 1p. sg. *-bi : pl. excl. *-ba :

incl. *-bidA, 2p. sg. *-ci : pl. *-ta. Generally, except in Dagur, the third person singular

and plural possessive suffixes have converged into the Common Mongolic shape *-ni.

Also, except in Dagur, the first person plural exclusive and inclusive forms have been

neutralized in favour of the exclusive form in the possessive paradigm, and in favour of

the inclusive form in the system of the predicative personal endings.

The demonstrative system in Proto-Mongolic was based on the two roots *e ‘this’and

*te ‘that’. These were probably never used alone, though the independent use of *te is

superficially suggested by synchronic data from Moghol and the languages of the Gansu-

Qinghai complex. In any case, the demonstrative roots were normally combined with

additional elements, both derivational and inflectional, yielding two sets of correlative

forms and derivatives. These, in turn, were closely parallelled by interrogative words

based on the root *ke ‘who, what’ (Table 1.5).

PROTO-MONGOLIC 19